事件驱动事件溯源架构

Architecture

事件溯源架构图，所有状态变更存储为不可变的领域事件序列，读取投影从事件流构建，快照优化用于快速聚合加载。该模板展示事件溯源如何通过保留每次状态转换的完整历史来消除数据丢失。对于需要完整审计跟踪、时间查询和事件重放能力的系统至关重要。

完整 FlowZap 代码

Command { # Command Side
n1: circle label:"User Command"
n2: rectangle label:"Validate Command"
n3: rectangle label:"Load Aggregate"
n4: rectangle label:"Apply Business Rules"
n5: rectangle label:"Emit Domain Events"
n1.handle(right) -> n2.handle(left)
n2.handle(right) -> n3.handle(left) [label="Valid"]
n3.handle(right) -> n4.handle(left) [label="Current State"]
n4.handle(right) -> n5.handle(left) [label="State Change"]
n5.handle(bottom) -> EventStore.n6.handle(top) [label="Append"]
}
EventStore { # Event Store
n6: rectangle label:"Append Event to Stream"
n7: rectangle label:"Assign Sequence Number"
n8: rectangle label:"Persist Immutable Event"
n9: rectangle label:"Publish to Projections"
n10: rectangle label:"Snapshot Manager"
n6.handle(right) -> n7.handle(left)
n7.handle(right) -> n8.handle(left) [label="Write-Ahead Log"]
n8.handle(right) -> n9.handle(left) [label="Notify"]
n9.handle(bottom) -> Projections.n11.handle(top) [label="Event Stream"]
n10.handle(top) -> Command.n3.handle(bottom) [label="Fast Load"]
}
Projections { # Read Projections
n11: rectangle label:"Event Handler"
n12: rectangle label:"Build Read Model"
n13: rectangle label:"Update Materialized View"
n14: rectangle label:"Serve Query"
n15: circle label:"Query Response"
n11.handle(right) -> n12.handle(left) [label="Transform"]
n12.handle(right) -> n13.handle(left) [label="Upsert"]
n13.handle(right) -> n14.handle(left) [label="Ready"]
n14.handle(right) -> n15.handle(left)
}
Replay { # Event Replay
n16: rectangle label:"Replay from Offset"
n17: rectangle label:"Rebuild Projection"
n18: rectangle label:"Verify Consistency"
n16.handle(right) -> n17.handle(left) [label="Historical Events"]
n17.handle(right) -> n18.handle(left) [label="Rebuilt"]
n18.handle(top) -> Projections.n13.handle(bottom) [label="Replace View"]
}

为什么需要这个工作流？

Traditional CRUD systems overwrite state on every update, losing the history of how the current state was reached. Event sourcing preserves every state change as an immutable event, enabling full audit trails, temporal queries, debugging by replaying events, and the ability to rebuild read models from scratch.

工作原理

Step 1: A user command is validated against the current aggregate state.
Step 2: The aggregate applies business rules and emits one or more domain events.
Step 3: Events are appended to the event store with a sequence number (never updated or deleted).
Step 4: Read projections consume the event stream and build materialized views.
Step 5: Snapshots are periodically created to speed up aggregate loading.
Step 6: Event replay can rebuild any projection from any point in time.

替代方案

Traditional CRUD with audit tables captures changes but cannot rebuild state. Change Data Capture provides similar benefits without changing application code. This template shows the full event sourcing architecture with projections and replay.

Key Facts

Template Name	事件驱动事件溯源架构
Category	Architecture
Steps	6 workflow steps
Format	FlowZap Code (.fz file)